From: Tantilov, E. S <emi...@in...> - 2007-04-26 13:38:57
|
Hi Torsten, =20 The ethtool stats you attached does not appear to be from the Intel = driver. Are you sure your problem is with the Intel NIC? =20 Thanks, Emil ________________________________ From: Torsten Kalix [mailto:Tor...@br...]=20 Sent: Wednesday, April 25, 2007 11:52 PM To: Tantilov, Emil S Cc: e10...@li... Subject: Re: [E1000-devel] Intel Server Board SE7320VP2 Thanks Emil for your answer. I am not using a "complete" server from Intel, but our local = hardware-vendor has made the server. Using a SE7320VP2-mainboard and and = SR4200-chassis. I now have tried the drivers 7.4.35, 7.3.20 and 7.0.41. = The old kernel 2.6.16 is working fine with driver 7.0.41 -- the new = kernel 2.6.20 is working for a few hrs before the intel-nic stops = responding, regardless of the driver version.=20 I also tried to look into the kernel-configuration but using a = distribution (Fedora Core 5) there are changes and kernel-patches where = I am just lost.=20 The only interesting thing I found (and attached) is the output from = "ethtool -S" while the problem exists. What does the value = "tx_mac_pause" mean ?? Thanks for your help, Torsten. Tantilov, Emil S wrote:=20 Torsten Kalix wrote: =20 First of all: Thanks for providing the driver ... I am using it on a =20 Thanks for using it and for reporting issues! =09 =20 couple of Linux systems, and it is running fine and stable. There are only two servers that have problems: =09 I have got two identical servers, using SE7320VP2. Running Fedora =20 Could you provide the exact model of the server you're using? =09 =20 Core 5 Kernel 2.6.16-1.2111_FC5smp and e1000 Version 7.0.41-NAPI. This configuration works fine. The computer has got a Marvell NIC and an Intel NIC: =09 05:03.0 Ethernet controller: Intel Corporation 82541GI/PI Gigabit Ethernet Controller (rev 05) Subsystem: Intel Corporation Unknown device 3466 Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr+ Stepping- SERR+ FastB2B- Status: Cap+ 66MHz+ UDF- FastB2B- ParErr- DEVSEL=3Dmedium =20 TAbort- <TAbort- <MAbort- >SERR- <PERR- =20 Latency: 64 (63750ns min), Cache Line Size 10 Interrupt: pin A routed to IRQ 21 Region 0: Memory at defa0000 (32-bit, non-prefetchable) [size=3D128K] Region 2: I/O ports at ee80 [size=3D64] Capabilities: [dc] Power Management version 2 Flags: PMEClk- DSI+ D1- D2- AuxCurrent=3D0mA PME(D0+,D1-,D2-,D3hot+,D3cold+) Status: D0 PME-Enable- DSel=3D0 DScale=3D1 PME- Capabilities: [e4] PCI-X non-bridge device Command: DPERE- ERO+ RBC=3D512 OST=3D1 Status: Dev=3D00:00.0 64bit- 133MHz- SCD- USC- = DC=3Dsimple DMMRBC=3D2048 DMOST=3D1 DMCRS=3D8 RSCEM- 266MHz- 533MHz- =09 =09 I tried to update our Kernel to 2.6.20 using the 2.6.20-1.2307.fc5smp from Fedora. I also updated the driver e1000 to Version 7.4.35-NAPI. =20 Have you tried drivers older than 7.4.35 (7.4.27, 7.3.20 etc)? You can get them from e1000.sf.net. It could be a regression problem. =09 =20 No other changes to hardware, hubs etc have been done. After some 1-2 hrs uptime the Intel Nic just stops responding. I can ping through the Marvell NIC, but not the Intel Nic. The machine does not crash, so local log-ons are still working. I viewed information from "dmesg", "ifconfig" and "ethtool", but without any hint about a problem (at least from what I know). =20 When it stops working check out stats using ethtool -S and send us the output. =09 =20 The resolution so far was re-boot and using the 2.6.16 kernel again, =20 Have you tried looking at differences between the configurations in the kernel config? =09 =20 but I would like to ask the everyone if one could give me a hint about troubleshooting or even a solution. =20 We'll try. =20 =20 Thanks for your help, =09 =09 =09 Torsten. =09 =20 =09 Thanks, Emil =09 = -------------------------------------------------------------------------= This SF.net email is sponsored by DB2 Express Download DB2 Express C - the FREE version of DB2 express and take control of your XML. No limits. Just data. Click to get it now. http://sourceforge.net/powerbar/db2/ _______________________________________________ E1000-devel mailing list E10...@li... https://lists.sourceforge.net/lists/listinfo/e1000-devel =09 =20 --=20 BREDEX GmbH Mauernstr. 33 38100 Braunschweig Tel.: +49-531-24330-0 Fax: +49-531-24330-99 http: www.bredex.de Gesch=E4ftsf=FChrer: Hans-J. Brede, Achim L=F6rke, Ulrich Obst Amtsgericht Braunschweig HRB 2450 |